A power line corridor photogrammetry method

Abstract

The application discloses a power line corridor oblique photogrammetry method and relates to the technical field of power transmission channel measurement. The specific steps are as follows: S1, a mixed flight platform is built, and aircraft and photographic equipment are matched based on the differences of the terrain and ground object features of different regions of the power line corridor. A fixed-wing unmanned aerial vehicle is used to collect basic images of the main body region of the corridor by carrying a downward-looking single lens, and a rotary-wing unmanned aerial vehicle is used to collect fine images of the building-dense or ground-object-rich region by carrying a multi-lens oblique camera, so as to build a strip mapping region framework of the power line corridor. The application realizes rapid rendering display of a mass model by using an efficient rendering engine in combination with distributed loading, memory and display memory optimization technology. The generated three-dimensional real scene result can realize efficient display of mass data and can be directly applied to various engineering application scenarios of the power line corridor, so as to provide high-quality three-dimensional data support for the planning, operation and management of the power line corridor.

Description

Technical Field

[0001] This invention relates to the field of power transmission channel measurement technology, specifically to a method for oblique photogrammetry of power line corridors. Background Technology

[0002] Power line corridors are characterized by complex terrain, irregular line distribution, and large spatial spans. Traditional photogrammetry methods are limited; relying solely on fixed-wing UAVs for data collection is insufficient for the detailed data acquisition needs of densely built-up areas with abundant features, while using only rotary-wing UAVs significantly increases flight costs. Furthermore, the lack of standardization in flight path planning and data processing makes it difficult to balance accuracy, efficiency, and cost in 3D reality data acquisition, hindering the efficient implementation of 3D reality data acquisition projects for power line corridors. To address this, we propose an oblique photogrammetry method for power line corridors. Summary of the Invention

[0003] To address the aforementioned technical problems, an oblique photogrammetry method for power line corridors is provided. This technical solution solves the problems of traditional photogrammetry methods in power line corridor applications, such as limited data acquisition methods, insufficient accuracy, high cost, and low efficiency.

[0004] To achieve the above objectives, the technical solution adopted by this invention is: a method for oblique photogrammetry of power line corridors, the specific steps of which are as follows: S1. Construct a hybrid flight platform and match aircraft and photography equipment according to the differences in terrain and land features in different areas of the power line corridor. Fixed-wing UAVs are equipped with downward-looking single lenses for basic image acquisition of the main area of ​​the corridor, while rotary-wing UAVs are equipped with multi-lens tilt cameras for fine image acquisition of areas with dense buildings or rich land features. This will construct a strip mapping area framework for the power line corridor. S2. For the strip mapping area, a single aerial photography zone is set up. A high-overlapping route laying strategy is adopted for route planning. The baseline and survey line extension design of the photography area boundary is carried out to ensure the integrity of data acquisition. Dual precision control is carried out based on CORS base station and ground control point for the entire process of image data acquisition. S3. The collected image data is preprocessed with DGPS positioning technology to solve the exterior orientation elements of the images; multiple iterative aerial triangulation based on the prior constraints of power tower coordinates is used to obtain high-precision aerial triangulation results; and oblique image modeling is performed in parallel with sub-units to obtain a three-dimensional model of the power line corridor. S4. Based on 3D models, it uses a high-efficiency rendering engine and combines distributed loading, memory optimization and video memory optimization technologies to render and display 3D scenes of massive model data. S5. Based on the visualization rendering results, perform lightweight noise reduction and manual repair of modeling defects in the 3D scene. Complete the accuracy verification through actual measurement checkpoints. Remodel unqualified areas until the accuracy requirements are met, and output a qualified 3D model. S6. Based on the output qualified 3D model, combined with the rendering engine, dynamic distributed loading, and memory and video memory optimization technology, output a complete set of surveying and mapping results and archive them, and connect to the power grid operation and maintenance business platform.

[0005] Preferably, the specific rules for differentiated matching in S1 are as follows: when the density of ground features in the area is lower than a preset threshold and the terrain is mainly plains or hills, a fixed-wing UAV is assigned to perform aerial photography tasks; when the density of buildings in the area exceeds a preset threshold or there is a dense area of ​​power equipment, a rotary-wing UAV is assigned to perform aerial photography tasks, and the aerial photography ranges of the two flight platforms maintain an overlap rate of more than 15% in the corridor overlap area for image data connection.

[0006] Preferably, the high overlap route laying strategy described in S2 has a forward overlap and a lateral overlap of 80%±5%, and dynamically compensates for the lateral overlap in sections where the terrain undulation exceeds the threshold based on the coarse DEM data of the strip mapping area. The greater the terrain undulation, the greater the lateral overlap compensation. The design of the extension of the baseline and survey line of the photographing area boundary is as follows: the forward boundary of the photographing area is extended outward by 20-25 baselines, and the lateral boundary is extended outward by 2-3 survey lines. The route is laid out in an east-west direction, and the first and last routes are extended outward by no less than three routes. The forward coverage exceeds the boundary of the photographing area by more than 30 baselines.

[0007] Preferably, the dual precision control based on CORS reference stations and ground control points described in S2 is as follows: existing CORS reference stations are reused as aerial photography base stations, the coordinates of the base station locations are provided by relevant parties and no ground connection is required, and the distance between the base station and the survey area does not exceed 50km; ground control points are set up on landmark features in the survey area, providing WGS84 ellipsoidal coordinates or local coordinates of the result coordinate system for geometric precision control and inspection.

[0008] Preferably, the specific conditions for the full-process image data acquisition described in S2 are as follows: the entire aerial photography flight is carried out using GPS navigation and strictly follows the planned data; the difference between the actual flight altitude and the designed flight altitude is less than 5% of the designed flight altitude; the operation is carried out on a clear day with horizontal visibility greater than 8 kilometers; the photography time is determined based on the solar altitude angle to ensure uniform illumination; and the POS or GPS raw data and control point coordinate data are recorded simultaneously.

[0009] Preferably, the multiple iterative aerial triangulation processing of the prior constraints on the power tower coordinates in S3 is as follows: the known coordinates of the power tower are obtained from the power grid management system, and they are introduced into the regional network adjustment equation as weighted soft constraints. Position constraints are applied to the image connection points where the power tower is located to reduce the accumulation of aerial triangulation errors in the images near the power tower. Feature point extraction, connection point matching, regional network adjustment and multiple iterative calculations are performed sequentially until convergence, generating high-precision aerial triangulation results.

[0010] Preferably, the parallel oblique image modeling process described in S3 performs suture line optimization planning when partitioning the model, so that the partition suture line avoids the areas where power lines and power towers are located, and guides the suture line to farmland, woodland or bare land without power facilities.

[0011] Preferably, the distributed loading technique in S4 is as follows: the 3D model is constructed into a multi-level LOD hierarchy by spatial block indexing, and only the model blocks within the viewport are loaded during rendering, while model blocks outside the viewport are released from video memory according to the most recently used strategy; the memory optimization adopts model data streaming loading, and the video memory optimization adopts a texture compression and geometry simplification hierarchical strategy.

[0012] Preferably, the accuracy verification through measured checkpoints described in S5 is as follows: checkpoints are evenly distributed within the strip mapping area, and the coordinates of the measured checkpoints are compared with the corresponding coordinates of the three-dimensional model. If the error in the plane position does not exceed 0.15 meters and the error in the elevation does not exceed 0.25 meters, it is considered qualified. After marking the unqualified areas, the model is re-flyed and remodeled according to the process from S1 to S3. The re-flying results are verified again until the accuracy requirements are met.

[0013] Preferably, the specific method for outputting the complete set of surveying and mapping results and archiving and connecting to the power grid operation and maintenance business platform as described in S6 is as follows: the qualified 3D model is spatially indexed and encoded according to the tower number and line section of the power line corridor, generating a complete set of surveying and mapping results package containing 3D model data, aerial triangulation results report, and accuracy verification record, which is then archived and stored according to a standardized directory structure; the surveying and mapping results package is pushed to the power grid operation and maintenance business platform through the API interface.

[0014] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention employs a hybrid flight platform combining fixed-wing and rotary-wing UAVs. Based on the differentiated terrain and feature characteristics of different regions, the aircraft and photographic equipment are matched accordingly. Fixed-wing UAVs perform large-scale, low-cost data collection in the main corridor area, while rotary-wing UAVs perform detailed data collection in complex areas, overcoming the limitations of traditional single-flight platforms that struggle to balance accuracy and cost. A single aerial photography zone is deployed to address the strip-like distribution characteristics of power line corridors, employing high-overlap flight paths and extending the boundaries of the photography zone to eliminate the risk of model loss due to insufficient edge overlap. Simultaneously, based on dual precision control using CORS base stations and ground control points, ground-based measurement from the base station is eliminated. Simultaneously, it achieves further control and inspection of geometric accuracy; it introduces multiple iterative aerial triangulation processing based on prior constraints of power tower coordinates to reduce the accumulation of aerial triangulation errors in images near power towers and improve the measurement accuracy of key facilities; it uses a parallel modeling module to process multiple model units simultaneously using a parallel processing hardware framework, significantly shortening the modeling runtime; it employs an efficient rendering engine combined with distributed loading, memory and video memory optimization techniques to achieve rapid rendering and display of massive models, and the generated 3D real-world results can achieve efficient display of massive data, which can be directly applied to various engineering application scenarios of power line corridors, providing high-quality 3D data support for the planning, operation, maintenance and management of power line corridors. Attached Figure Description

[0015] Figure 1 This is a flowchart of the steps of the present invention. Detailed Implementation

[0016] The following description is intended to disclose the invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art.

[0017] A method for oblique photogrammetry of power line corridors includes the following steps: S1: Build a hybrid flight platform and match aircraft and photography equipment according to the differences in terrain and feature characteristics of different areas of the power line corridor. Fixed-wing UAVs are equipped with downward-looking single lenses for basic image acquisition of the main area of ​​the corridor, while rotary-wing UAVs are equipped with multi-lens tilt cameras for fine image acquisition of areas with dense buildings or rich features. This will construct a strip mapping area framework for the power line corridor.

[0018] The specific rules for differentiated matching are as follows: when the density of ground features in the area is lower than a preset threshold and the terrain is mainly plains or hills, a fixed-wing UAV is assigned to perform aerial photography tasks; when the building density in the area exceeds a preset threshold or there are areas with dense power equipment, a rotary-wing UAV is assigned to perform aerial photography tasks. The aerial photography ranges of the two flight platforms maintain an overlap rate of more than 15% in the corridor overlap area for image data stitching. The fixed-wing UAV completes lens focal length and exposure parameter calibration to ensure that the equipment can achieve basic image acquisition with a high degree of overlap; the rotary-wing UAV completes synchronization calibration and tilt angle adjustment of each lens to ensure that detailed tilt images from multiple perspectives can be acquired in areas with dense buildings or abundant ground features; at the same time, the GPS positioning modules and flight control systems of the two UAVs are debugged to ensure the accuracy of the flight trajectory.

[0019] S2: For strip mapping areas, a single aerial photography zone is set up, and a high-overlap route laying strategy is adopted for route planning. The baseline and survey line extension design of the photography area boundary is carried out to ensure the integrity of data acquisition. Based on CORS base station and ground control point, dual precision control is carried out for the entire process of image data acquisition.

[0020] The high-overlap route laying strategy uses a forward overlap and lateral overlap of 80% ± 5%, and dynamically compensates for lateral overlap in sections with terrain undulations exceeding a threshold based on coarse DEM data of the strip mapping area. The greater the terrain undulation, the greater the lateral overlap compensation, which is determined by the following formula: in, The compensated lateral overlap, The baseline lateral overlap is 80%. This is the terrain compensation coefficient (value range: 0.5-1.0). Due to local topographical elevation differences, For flight design purposes, this compensation mechanism ensures that the actual overlap caused by terrain undulations in mountain corridors is no less than the lower limit of 75%, eliminating insufficient image overlap caused by elevation differences.

[0021] The design for extending the baselines and survey lines of the photographing area boundary is as follows: the boundary of the photographing area in the heading direction is extended outward by 20-25 baselines, and the boundary in the lateral direction is extended outward by 2-3 survey lines. The overall flight path is laid out in an east-west direction, with the first and last flight paths extending outward by no less than three flight paths, and the heading coverage exceeding the boundary of the photographing area by more than 30 baselines.

[0022] Dual precision control based on CORS base stations and ground control points is as follows: existing CORS base stations are reused as aerial survey base stations, the coordinates of the base station locations are provided by relevant parties and no ground connection is required, and the distance between the base station and the survey area does not exceed 50km; ground control points are set up on landmark features in the survey area, providing WGS84 ellipsoidal coordinates or local coordinates of the result coordinate system for geometric precision control and inspection.

[0023] The specific conditions for the full-process image data acquisition are as follows: the entire aerial photography flight is guided by GPS and executed strictly according to the planned data; the difference between the actual flight altitude and the designed flight altitude is less than 5% of the designed flight altitude; the operation is carried out on a clear day with horizontal visibility greater than 8 kilometers; the photography time is determined based on the solar altitude angle to ensure uniform illumination; and the POS or GPS raw data and control point coordinate data are recorded simultaneously.

[0024] S3: The acquired image data is sequentially preprocessed using DGPS positioning technology to solve the exterior orientation elements of the images; multiple iterative aerial triangulation processes based on the prior constraints of power tower coordinates are used to obtain high-precision aerial triangulation results; and oblique image modeling processing in parallel with sub-units is used to obtain a three-dimensional model of the power line corridor.

[0025] The trajectory data preprocessing is as follows: the collected POS or GPS raw data is imported into dedicated post-processing software, and the exterior orientation elements of each image at the exposure time are precisely calculated according to carrier phase measurement differential GPS (DGPS) positioning technology; the quality of the calculated data is checked to ensure that the data is complete, uninterrupted, and of qualified quality; finally, through parameter conversion and image number correspondence, the exterior orientation elements of each aerial image are output as the basic data for subsequent processing.

[0026] The iterative aerial triangulation process, incorporating prior constraints on power tower coordinates, involves obtaining the known coordinates of the power towers from the power grid management system and introducing them as weighted soft constraints into the regional network adjustment equations. The observation equations for the regional network adjustment are: in, Let be the residual vector of the observed values. To design the matrix, Correction vector for unknown parameters, Let be the observed value vector. After introducing the prior constraint of the power tower coordinates, add constraint terms to the normal equations. The constraint conditions are as follows: in, For the correction of the power tower location parameters, Given the coordinate vector of the power tower, This is the constraint weight matrix. The constraint weights are determined by the following formula: in, The weight of a single power tower constraint point. The variance is the unit weight. Let V be the prior variance of the known coordinates of the power tower. The value is determined based on the accuracy level of the coordinates in the power grid management system. For power tower coordinates obtained by GPS measurement, Use a range of 0.05-0.10 meters; for coordinates obtained through RTK measurement, Take 0.02-0.05 meters. The larger the value, the stronger the constraint, and the smaller the cumulative aerial triangulation error of the image near the power tower. Feature point extraction, tie point matching, regional network adjustment, and multiple iterative calculations are performed sequentially until convergence, generating high-precision aerial triangulation results.

[0027] The parallel oblique image modeling process involves: constructing a modeling workstation with a CPU-based hardware architecture capable of parallel processing; importing the aerial triangulation results from aerial triangulation into the modeling software; dividing each encrypted partition of the image area into multiple independent model units; and utilizing the workstation's parallel processing capabilities to simultaneously perform 3D modeling on each model unit. During model partitioning, suture line optimization planning is performed, with the suture line cost function defined as: in, For the first The cost of color difference at each candidate suture location Penalty factor for avoiding power facilities (value ranges from 10 to 100). For indicator functions, when the first The candidate locations fall within the buffer zone of power lines or power towers. The value is 1 if the condition is met, and 0 otherwise. By minimizing the cost function of the suture line, the suture line is guided to avoid areas where power lines and towers are located, and to farmland, woodland, or bare land without power facilities, thus preventing linear features from breaking across the suture line and causing missing power line models. The overall 3D model of the power line corridor is then completed.

[0028] S4: Based on 3D models, it uses a high-efficiency rendering engine and combines distributed loading, memory optimization and video memory optimization technologies to render and display 3D scenes of massive model data.

[0029] Distributed loading technology involves constructing a multi-level LOD hierarchy by spatially partitioning and indexing the 3D model. The LOD level is determined by the following formula: in, For LOD levels, The maximum spatial span of the model, This represents the distance between the current viewpoint and the center of the model block. During rendering, only model blocks within the viewport and matching the current view distance at the corresponding LOD level are loaded. Model blocks outside the viewport are released from video memory according to the most recently used strategy. Memory optimization employs streaming loading of model data, while video memory optimization uses a hierarchical strategy of texture compression and geometry simplification. A self-developed high-efficiency rendering engine is used, equipped with dedicated tools specific to the power industry. This engine can efficiently organize and process 3D model materials, enabling rapid overlay and loading of massive amounts of tower and corridor terrain and feature model data. This completes the rendering of the 3D scene of the power line corridor, achieving a visual representation of the 3D reality results.

[0030] S5: Based on the visualization rendering results, perform lightweight noise reduction and manual repair of modeling defects in the 3D scene, complete the accuracy verification through actual measurement checkpoints, remodel unqualified areas until the accuracy requirements are met, and output a qualified 3D model.

[0031] Accuracy verification is accomplished through actual measurement of checkpoints: checkpoints are evenly distributed within the strip mapping area, and the coordinates of the actual checkpoints are compared with the corresponding coordinates of the 3D model. The mean square error of the planar position and the mean square error of the elevation are calculated by the following formulas: in, For the number of checkpoints, For the first The coordinates of each checkpoint in the 3D model For the first The measured coordinates of each checkpoint. A horizontal position error of no more than 0.15 meters and an elevation error of no more than 0.25 meters are considered acceptable. Unacceptable areas are marked and then re-fly and modeled according to the S1 to S3 process. The re-fly results are then verified again until the accuracy requirements are met.

[0032] S6: Based on the output qualified 3D model, combined with rendering engine, dynamic distributed loading, memory and video memory optimization technology, outputs a complete set of surveying and mapping results and archives them, and connects to the power grid operation and maintenance business platform.

[0033] The specific method for outputting and archiving the complete set of surveying and mapping results and connecting them to the power grid operation and maintenance business platform is as follows: Qualified 3D models are spatially indexed and encoded according to the tower numbers and line sections of the power line corridor, generating a complete set of surveying and mapping results package containing 3D model data, aerial triangulation results report, and accuracy verification records, which is then archived and stored according to a standardized directory structure; the surveying and mapping results package is pushed to the power grid operation and maintenance business platform through the API interface to achieve data exchange and visualization call between 3D real-scene data and business scenarios such as transmission channel inspection, hidden danger investigation, and channel management.

[0034] Example 1 The project focused on a power line corridor in a plain-town transition zone. The corridor is 30 km long, with approximately 22 km in the plain and approximately 8 km in the town. The plain section has a terrain undulation of less than 10 meters, and the terrain features are mainly farmland and bare land. The town section has a higher building density and several areas with dense power equipment.

[0035] S1: Construct a hybrid flight platform. For the plains section, a fixed-wing UAV equipped with a single downward-looking lens (35mm focal length) will be used, with a flight altitude set at 300 meters. For the urban section, a rotary-wing UAV equipped with a 5-lens tilting camera (25mm downward-looking focal length, 35mm tilting focal length) will be used, with a flight altitude set at 200 meters. The two flight platforms will maintain a 20% aerial photography overlap in the plains-urban transition zone to ensure seamless image data integration. After equipment debugging and calibration, a strip-shaped mapping area framework will be constructed.

[0036] S2: One aerial survey zone is established, with flight path planning using 80% forward overlap and 80% lateral overlap. The plains section has minimal terrain undulation, requiring no lateral overlap compensation; the urban section contains building clusters with elevation differences of 5-15 meters, and compensation is based on the terrain undulation formula. , rice, The lateral overlap was increased to 85.4% after compensation. The forward boundary of the survey area was extended outward by 22 baselines, and each lateral boundary was extended outward by 2 survey lines. The initial and final flight paths were extended outward by 3 flight paths, resulting in forward coverage exceeding 30 baselines beyond the survey area's boundary. An existing CORS reference station 35 km from the survey area was reused, eliminating the need for ground-based measurements. Twelve ground control points were established on landmark features within the survey area to provide WGS84 ellipsoidal coordinates. On clear days with horizontal visibility greater than 8 km, the photography times were determined based on the solar altitude angle: 9:00-11:00 AM and 2:00-4:00 PM. The difference between the actual flight altitude and the designed flight altitude was less than 3%. Raw POS data and control point coordinate data were recorded simultaneously.

[0037] S3: Import the raw POS data into the post-processing software, calculate the exterior orientation elements for each aerial photograph's exposure time using DGPS positioning technology, and verify the calculation quality. The known coordinates of 18 power towers (obtained via GPS measurement) are incorporated into the aerial triangulation process. A weighted soft constraint (0.08 meters) is used in the regional network adjustment. Feature point extraction, connector matching, and regional network adjustment are performed sequentially, with convergence after three iterations. Modeling is performed using an 8-core parallel CPU framework, dividing the camera region into six model units for parallel modeling. The stitching cost function... Taking 50, the suture successfully avoided all 18 power towers and power line areas, guiding them to farmland and bare land.

[0038] S4: Employs a high-efficiency rendering engine to construct a 4-level LOD hierarchy for 3D models, achieving a rendering frame rate of 45fps, enabling smooth 3D scene rendering and display of massive model data.

[0039] S5: 20 checkpoints were set up for accuracy verification. The mean square error of the horizontal position was 0.098 meters and the mean square error of the elevation was 0.172 meters, both of which met the accuracy requirements. No re-flying was required, and a qualified 3D model was output.

[0040] S6: Spatial index coding is performed according to tower number and line section, surveying and mapping result package is generated and archived, and pushed to the power grid operation and maintenance business platform through API interface.

[0041] Example 2 The project focused on a power line corridor in a mountainous area. The corridor is 25km long and the terrain is mainly mountainous, with an elevation range of 300-1200 meters. The terrain is highly undulating, with some sections having an elevation difference of more than 200 meters. The terrain features mainly consist of woodlands and bare rocks, with a few villages in some areas.

[0042] S1: Fixed-wing UAVs equipped with a single downward-looking lens (35mm focal length) are assigned to the main mountainous section, with a flight altitude set at 400 meters; rotary-wing UAVs equipped with a 5-lens tilt camera (25mm downward-looking focal length, 35mm tilt focal length) are assigned to the villages and areas with dense power equipment, with a flight altitude set at 250 meters. The two flight platforms maintain an 18% aerial photography overlap rate during the transition section.

[0043] S2: One aerial survey zone is established, with flight path planning using 82% forward overlap and 82% lateral overlap. For mountainous areas with significant elevation differences, the terrain undulation compensation formula is used. , rice, After compensation, the lateral overlap was increased to 89.2%, ensuring sufficient overlap in mountainous imagery. The forward boundary of the survey area was extended outward by 25 baselines, and each lateral boundary was extended outward by 3 survey lines. The initial and final flight paths were extended outward by 4 flight paths, resulting in forward coverage exceeding the survey area's boundary by 35 baselines. An existing CORS reference station 42 km from the survey area was reused, and 15 ground control points were established. Operations were conducted on clear days with horizontal visibility greater than 8 km, and the difference between the actual flight altitude and the design flight altitude was less than 4%.

[0044] S3: The preprocessing and calculation of the exterior orientation elements of the DGPS trajectory data are qualified. Aerial triangulation processing incorporates the known coordinates of 15 power towers (obtained through RTK measurement). A weighted soft constraint (0.03 meters) was introduced into the regional network adjustment, and convergence was achieved after 4 iterations. The modeling process divided the camera region into 8 model units for parallel modeling, and the stitching cost function... Taking 80, the suture successfully avoided all 15 power towers and power line areas, guiding them to woodland and bare rock areas.

[0045] S4: Employs a high-efficiency rendering engine, constructs a 5-level LOD hierarchy, and achieves a rendering frame rate of 38fps.

[0046] S5: 18 checkpoints are set up, with a mean error of 0.112 meters in planar position and 0.201 meters in elevation, meeting the accuracy requirements and outputting a qualified 3D model.

[0047] S6: Spatial index coding is performed according to tower number and line section, surveying and mapping result package is generated and archived, and pushed to the power grid operation and maintenance business platform through API interface.

[0048] Example 3 The project focuses on a power line corridor with a large span and complex terrain. The corridor is 50km long and passes through plains, hills and mountains. The plain section is about 15km long, the hill section is about 18km long, the mountain section is about 17km long, and the urban section is about 5km long. The terrain features are complex and diverse.

[0049] S1: Fixed-wing UAVs equipped with a downward-looking single-lens camera (35mm focal length) are assigned to the plains and hilly sections, with a flight altitude set at 350 meters; rotary-wing UAVs equipped with a 5-lens tilting camera are assigned to the urban sections and some mountainous areas with dense power equipment, with a flight altitude set at 220 meters. All three aerial survey sections maintain an aerial survey overlap rate of greater than 15% in the transition sections.

[0050] S2: One aerial survey zone is established, with flight path planning using 83% forward overlap and 83% lateral overlap. No compensation is required in the plains section; in the hilly section, the local elevation difference is 50-80 meters, and [the following parameters are used]. , After compensation, the lateral overlap increased to 86.3%; in mountainous areas with an elevation difference of 150-300 meters, [the following parameters were taken]. , After compensation, the lateral overlap was increased to 93.5%. The forward boundary of the survey area was extended outward by 25 baselines, and the lateral boundaries were extended outward by 3 survey lines on each side. The initial and final flight paths were extended outward by 4 flight paths, resulting in forward coverage exceeding 32 baselines beyond the survey area's boundary. An existing CORS reference station 48 km from the survey area was reused, and 22 ground control points were established. Operations were conducted on clear days with horizontal visibility greater than 8 km, and the difference between the actual flight altitude and the design flight altitude was less than 4.5%.

[0051] S3: The preprocessing and calculation of the exterior orientation elements of the DGPS trajectory data are qualified. Aerial triangulation incorporates the known coordinates of 32 power towers, of which 18 were obtained from GPS measurements. Meters, 14 bases obtained by RTK measurement The meter, as a weighted soft constraint, is introduced into the regional network adjustment, and convergence is achieved after 5 iterations. The modeling process divides the camera region into 12 model units for parallel modeling, and the stitching cost function... Taking 60, the suture successfully avoided all 32 power towers and power line areas.

[0052] S4: Employs a high-efficiency rendering engine, constructs a 5-level LOD hierarchy, and achieves a rendering frame rate of 35fps.

[0053] S5: 25 checkpoints were set up. Initial verification found that the planar mean square error of 0.18 meters in two areas of the mountain section did not meet the accuracy requirements. The two areas were re-modeled by flight testing. After the flight testing, the planar mean square errors were 0.098 meters and 0.105 meters respectively, which met the accuracy requirements. The qualified 3D model was output.

[0054] S6: Spatial index coding is performed according to tower number and line section, surveying and mapping result package is generated and archived, and pushed to the power grid operation and maintenance business platform through API interface.

[0055] Comparative Example 1 Comparative Example 1 uses a single fixed-wing UAV equipped with a downward-looking single lens to conduct aerial photography of the same corridor as in Example 1, with all other parameters being the same as in Example 1.

[0056] A single fixed-wing UAV cannot acquire detailed oblique images of densely built-up urban areas. The missing texture rate of building side surfaces in the 3D model of urban areas is as high as 35%, and the fine structure of power equipment cannot be restored. In contrast, the missing texture rate of building side surfaces in the hybrid flight platform solution of Example 1 is only 2%, and the fine structure of power equipment is fully restored.

[0057] Comparative Example 2 Comparative Example 2 uses conventional aerial triangulation processing (without introducing prior constraints on power tower coordinates) to process the same corridor data from Example 1, with the remaining parameters being consistent with Example 1.

[0058] Conventional aerial triangulation methods show significant error accumulation in areas with dense power towers, with tower position deviations reaching 0.25-0.45 meters. However, after introducing prior constraints on power tower coordinates in Example 1, the tower position deviations are controlled within 0.05-0.12 meters, improving accuracy by approximately 70%.

[0059] Comparative Example 3 Comparative Example 3 uses conventional uniform partitioning modeling (without suture line optimization planning) to process the same corridor data from Example 2, with the remaining parameters being the same as in Example 1.

[0060] The conventional uniform partitioning resulted in three suture lines crossing the power line area, causing the power line model to break and be missing at the corresponding locations, with a power line model integrity rate of only 82%. However, after the suture line optimization planning in Example 2, all suture lines avoided the power facility area, and the power line model integrity rate reached 100%.

[0061] Performance Comparison The above performance comparison shows that: Compared to a single fixed-wing drone, the hybrid flight platform significantly improves the integrity of building side textures and the completeness of power line models. Although the flight cost increases, it is far lower than that of a single rotor drone solution (the relative cost of a single rotor solution is about 2.5-3.0). Compared to conventional aerial triangulation, the prior constraint aerial triangulation method for power tower coordinates improves the positional accuracy of power towers by approximately 70% and reduces the elevation error by approximately 30%. Compared to conventional uniform zoning, the suture line optimization planning improved the power line model integrity rate from 82% to 100%, eliminating the problem of linear features breaking across the suture line.

[0062] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention.

Claims

1. A method for oblique photogrammetry of power line corridors, characterized in that, The specific steps are as follows: S1. Construct a hybrid flight platform and match aircraft and photography equipment according to the differences in terrain and land features in different areas of the power line corridor. Fixed-wing UAVs are equipped with downward-looking single lenses for basic image acquisition of the main area of ​​the corridor, while rotary-wing UAVs are equipped with multi-lens tilt cameras for fine image acquisition of areas with dense buildings or rich land features. This will construct a strip mapping area framework for the power line corridor. S2. For the strip mapping area, a single aerial photography zone is set up. A high-overlapping route laying strategy is adopted for route planning. The baseline and survey line extension design of the photography area boundary is carried out to ensure the integrity of data acquisition. Dual precision control is carried out based on CORS base station and ground control point for the entire process of image data acquisition. S3. The collected image data is preprocessed with DGPS positioning technology to solve the exterior orientation elements of the images; multiple iterative aerial triangulation based on the prior constraints of power tower coordinates is used to obtain high-precision aerial triangulation results; and oblique image modeling is performed in parallel with sub-units to obtain a three-dimensional model of the power line corridor. S4. Based on 3D models, it uses a high-efficiency rendering engine and combines distributed loading, memory optimization and video memory optimization technologies to render and display 3D scenes of massive model data. S5. Based on the visualization rendering results, perform lightweight noise reduction and manual repair of modeling defects in the 3D scene. Complete the accuracy verification through actual measurement checkpoints. Remodel unqualified areas until the accuracy requirements are met, and output a qualified 3D model. S6. Based on the output qualified 3D model, combined with the rendering engine, dynamic distributed loading, and memory and video memory optimization technology, output a complete set of surveying and mapping results and archive them, and connect to the power grid operation and maintenance business platform.

2. The method for oblique photogrammetry of power line corridors according to claim 1, characterized in that, The specific rules for differentiated matching described in S1 are as follows: when the density of ground features in the area is lower than a preset threshold and the terrain is mainly plains or hills, a fixed-wing UAV is assigned to perform aerial photography tasks; when the density of buildings in the area exceeds a preset threshold or there is a dense area of ​​power equipment, a rotary-wing UAV is assigned to perform aerial photography tasks. The aerial photography ranges of the two flight platforms maintain an overlap rate of more than 15% in the corridor overlap area for image data connection.

3. The method for oblique photogrammetry of power line corridors according to claim 1, characterized in that: The high-overlap route laying strategy described in S2 has a forward overlap and a lateral overlap of 80% ± 5%. Based on the coarse DEM data of the strip mapping area, the lateral overlap is dynamically compensated for sections with terrain undulations exceeding the threshold. The greater the terrain undulation, the greater the lateral overlap compensation. The design of the baseline and survey line extension of the survey area boundary is as follows: the forward boundary of the survey area is extended outward by 20-25 baselines, and the lateral boundary is extended outward by 2-3 survey lines. The route is laid out in an east-west direction, with the first and last routes extended outward by no less than three routes, and the forward coverage exceeds the boundary of the survey area by more than 30 baselines.

4. The method for oblique photogrammetry of power line corridors according to claim 1, characterized in that, The dual precision control based on CORS reference stations and ground control points described in S2 is as follows: existing CORS reference stations are reused as aerial photography base stations, the coordinates of the base station locations are provided by relevant parties and no ground connection is required, and the distance between the base station and the survey area does not exceed 50km; ground control points are set up on landmark features in the survey area, providing WGS84 ellipsoidal coordinates or local coordinates of the result coordinate system for geometric precision control and inspection.

5. The method for oblique photogrammetry of power line corridors according to claim 1, characterized in that, The specific conditions for the full-process image data acquisition described in S2 are as follows: the entire aerial photography flight is carried out using GPS navigation and strictly follows the planned data; the difference between the actual flight altitude and the designed flight altitude is less than 5% of the designed flight altitude; the operation is carried out on a clear day with horizontal visibility greater than 8 kilometers; the photography time is determined based on the solar altitude angle to ensure uniform illumination; and the POS or GPS raw data and control point coordinate data are recorded simultaneously.

6. The method for oblique photogrammetry of power line corridors according to claim 1, characterized in that, The multiple iterative aerial triangulation process for the prior constraints of the power tower coordinates described in S3 is as follows: the known coordinates of the power tower are obtained from the power grid management system and introduced into the regional network adjustment equation as a weighted soft constraint condition. Position constraints are applied to the image connection points where the power tower is located to reduce the accumulation of aerial triangulation errors in the images near the power tower. Feature point extraction, connection point matching, regional network adjustment and multiple iterative calculations are performed sequentially until convergence is achieved, generating high-precision aerial triangulation results.

7. The method for oblique photogrammetry of power line corridors according to claim 1, characterized in that: The parallel oblique image modeling process described in S3 performs suture line optimization planning when partitioning the model, so that the partition suture line avoids the areas where power lines and power towers are located, and guides the suture line to farmland, woodland or bare land without power facilities.

8. The method for oblique photogrammetry of power line corridors according to claim 1, characterized in that, The distributed loading technique described in S4 is as follows: the 3D model is constructed into a multi-level LOD hierarchy by spatial block indexing, and only the model blocks within the viewport are loaded during rendering. Model blocks outside the viewport are released from video memory according to the most recently used strategy. The memory optimization adopts model data streaming loading, and the video memory optimization adopts a texture compression and geometry simplification hierarchical strategy.

9. The method for oblique photogrammetry of power line corridors according to claim 1, characterized in that, The accuracy verification through actual measurement checkpoints described in S5 is as follows: checkpoints are evenly distributed within the strip mapping area, and the coordinates of the actual measurement checkpoints are compared with the corresponding coordinates of the three-dimensional model. If the error in the plane position does not exceed 0.15 meters and the error in the elevation does not exceed 0.25 meters, it is considered qualified. After marking the unqualified areas, the model is re-flyed and remodeled according to the process from S1 to S3. The results of the re-flying are verified again until the accuracy requirements are met.

10. The method for oblique photogrammetry of power line corridors according to claim 1, characterized in that, The specific method for outputting and archiving the complete set of surveying and mapping results and connecting to the power grid operation and maintenance business platform as described in S6 is as follows: the qualified 3D model is spatially indexed and encoded according to the tower number and line section of the power line corridor, generating a complete set of surveying and mapping results package containing 3D model data, aerial triangulation results report, and accuracy verification record, which is then archived and stored according to a standardized directory structure; the surveying and mapping results package is pushed to the power grid operation and maintenance business platform through the API interface.

Images